286 beport— 1879. 



stout plate of glass about 12 mm. thick, and pressing the melted lac into a disc 

 hy another similar plate of glass, the two plates being separated by slips of thin 

 glass. 



The plate so pressed was cut into small pieces, and the thickness of each piece 

 measured, the fragments having the same thickness (within ^~ of an inch) beino- 

 placed together. 



A stout piece of good plate glass was covered with tin-foil, and on this in 

 proper position three little pieces of lac were laid to support the first small circle of 

 glass ; on this three other pieces were placed over the first three pieces to support 

 the first large circle, and so on to the end. Care was taken to select the three 

 pieces supporting each plate of the same thickness. 



The mean distance between the plates was determined by measuring the height 

 of the pile when completed, and the height after removing the shellac separating 

 the plates. 



The mean area of the shellac by weighing the fragments used and comparing 

 their weight with that of a piece of the plate from which they were cut of such size 

 that its dimensions could be measured with some accuracy. 



The three other condensers used (called A, B, C) were 



A. The condenser referred to in Dr. Muirhead's paper. 



B. A condenser made by Messrs. Warden & Clark at some time not known by 



me and used for comparison. 



C. A small mica condenser, capacity about 0-1 mfd. made in 1867, and copied 



from the condenser described by Professor Jenkin in his paper ' British 

 Association Reports,' 1867. 

 The method of observation was as follows : — 



a. The time of oscillation of a galvanometer needle was observed. 



b. The deflection of the galvanometer needle produced by a steady current was 

 noted, the current being either that produced by the battery employed to charge 

 the condenser flowing through a resistance of from 500,000 to 1,000,000 ohms or 

 else a current produced by any other battery, the electromotive force between 

 two points in the circuit separated by a known resistance being determined 

 in terms of the electromotive force employed to charge the condenser. 



c. The condenser was charged and discharged several times. As it was found 

 impracticable to maintain the galvanometer needle at rest absolutely when the 

 time of oscillation was great, the method employed was to reduce the oscillations 

 to a small amount, 2, 3, or 4 divisions, read three successive oscillations and 

 discharge the condenser at the moment that the needle was at rest and on the 

 point of changing its direction of motion. The mean of the second deflection and 

 the half sum of the first and third deflections is taken as the zero, and the total 

 excursion of the needle in the case in question is not altered by the fact that the 

 needle started from a position of rest but not of equilibrium. 



In fact, let G be the angle defining the extreme excursion of the needle due 

 a given impulse applied to it when it was at rest in a position of equilibrium 



6 the extreme excursion when the impulse is applied, when the position of the 

 needle is defined by the angle 6, and is oscillating through an angle <9„ on either side 

 of its zero point. Then (neglecting the effect of air resistance) 



and, therefore, if 8 X =<9 2 , or the impulse is applied at the moment the needle is at 

 the extremity of an excursion, 



- e(\-h 6 I + &c.) 



= 6 if§ S 

 6 

 may be neglected, as it may in all the cases occurring in these experiments. 



